This survey summarizes the state-of-the-art research on network coding, mainly focusing on its applications to computer networking. Network coding generalizes traditional store-and-forward routing techniques by allowing intermediate nodes in networks to encode several received packets into a single coded packet before forwarding. Network coding was proposed in 2000, and since then, it has been studied extensively in the field of computer networking. In this survey, we first summarize linear network coding and provide a taxonomy of network coding research, i.e., the network coding design problem and network coding applications. Moreover, the latter is subdivided into throughput/capacity enhancement, robustness enhancement, network tomography, and security. We then discuss the fundamental characteristics of network coding and diverse applications of network coding in details, following the above taxonomy.

In this paper, we consider network monitoring techniques to estimate communication qualities in wide-area mobile networks, where an enormous number of heterogeneous components such as base stations, routers, and servers are deployed. We assume that average delays of neighboring base stations are comparable, most of servers have small delays, and delays at core routers are negligible. Under these assumptions, we propose Heterogeneous Delay Tomography (HDT) to estimate the average delay at each network component from end-to-end round trip times (RTTs) between mobile terminals and servers. HDT employs a crowdsourcing approach to collecting RTTs, where voluntary mobile users report their empirical RTTs to a data collection center. From the collected RTTs, HDT estimates average delays at base stations in the Graph Fourier Transform (GFT) domain and average delays at servers, by means of Compressed Sensing (CS). In the crowdsourcing approach, the performance of HDT may be degraded when the voluntary mobile users are unevenly distributed. To resolve this problem, we further extend HDT by considering the number of voluntary mobile users. With simulation experiments, we evaluate the performance of HDT.

We propose a home base-aware store-carry-forward routing scheme using location-dependent utilities of nodes, which adopts different message forwarding strategies depending on location where nodes encounter. Our routing scheme first attempts to deliver messages to its home base, the area with the highest potential for the presence of the destination node in the near future. Once a message copy reaches its home base, message dissemination is limited within the home base, and nodes with message copies wait for encountering the destination node. To realize our routing scheme, we use two different utilities of nodes depending on location: Outside the home base of a message, nodes approaching to the home base have high utility values, while within the home base, nodes staying the home base have high utility values. By using these utilities properly, nodes with message copies will catch the destination node “by ambush” in the home base of the destination node. Through simulation experiments, we demonstrate the effectiveness of our routing scheme.

Recent improvements in the performance of end-computers and networks have made it feasible to construct a grid system over the Internet. A grid environment consists of many computers, each having a set of components and a distinct performance. These computers are shared among many users and managed in a distributed manner. Thus, it is important to focus on a situation in which the computers are used unevenly due to decentralized management by different task schedulers. In this study, which is a preliminary investigation of the performance of task allocation schemes employed in a decentralized environment, the average execution time of a long-lived task is analytically derived using the M/G/1-PS queue. Furthermore, assuming a more realistic condition, we evaluate the performance of some task allocation schemes adopted in the analysis, and clarify which scheme is applicable to a realistic grid environment.

The layer3 switch enables us to fast transmit IP datagrams using the cut-through technique. The current layer3 router would become bottleneck in terms of delay performance as the amount of traffic injected into high speed networks gets relatively large. Thus, the layer3 switch should be an important element constructing the next generation Internet backbone. In this paper, we analyze the cut-through rate, the datagram waiting time and the mis-ordered rate of a layer3 switch in case of flow-driven connection setup. In the analysis, by using 3-state Markov modulated Bernoulli process (MMBP), we model the arrival process of IP flow and IP datagram from each source. Furthermore, we investigate impacts of the arrival rate and the average datagram length on the performance.

In future high-speed networks, provision of diverse multimedia services with strict quality-of-service (QoS) requirements, such as bandwidth, delay and so on, is desired. QoS routing is a possible solution to handle these services. Generally, a path selection for QoS routing is formulated as a shortest path problem subject to multiple constraints. However, it is known to be NP-complete when more than one QoS constraint is imposed. As a result, many heuristic algorithms have been proposed so far. The authors proposed a path selection algorithm Fallback+ for QoS routing, which focuses not only on the path selection with multiple constraints but also on the efficient use of network resources. This paper proposes an enhanced version of Fallback+, named Enhanced Fallback+, where in a shrewd way, it keeps tentative paths produced in the conventional Fallback algorithm with Dijkstra's algorithm. Simulation experiments prove the excellent performance of Enhanced Fallback+, compared with the original Fallback+ and other existing path selection algorithms.

This paper studies the local Poisson property of aggregated IP traffic. First, it describes the scenario where IP traffic presents a Poisson-like characteristic within some limited range of time scales when packets from independent traffic streams are aggregated. Each of the independent traffic streams corresponds to a series of correlated IP packets such as those of a transport connection. Since the Poisson-like characteristic is observed only within some limited range of time scales, we call this characteristic the local Poisson property. The limited range of time scales of the local Poisson property can be estimated from a network configuration and characteristics of transport connections. Second, based on these observations, we seek the possibility to apply an ordinary Poisson process to evaluation of the packet loss probability in IP networks. The analytical investigation, where IP traffic is modeled by a superposition of independent branching Poisson processes that presents the local Poisson property, suggests that the packet loss probability can be estimated by a finite-buffer queue with a Poisson process when the buffer size is within a certain range. The investigation is verified by simulations. These findings expand the applicability of conventional Poisson-based approaches to IP network design issues.

In the IEEE 802.11 MAC protocol, access points (APs) are given the same priority as wireless terminals in terms of acquiring the wireless link, even though they aggregate several downlink flows. This feature leads to a serious throughput degradation of downlink flows, compared with uplink flows. In this paper, we propose a dynamic contention window control scheme for the IEEE 802.11e EDCA-based wireless LANs, in order to achieve fairness between uplink and downlink TCP flows while guaranteeing QoS requirements for real-time traffic. The proposed scheme first determines the minimum contention window size in the best-effort access category at APs, based on the number of TCP flows. It then determines the minimum and maximum contention window sizes in higher priority access categories, such as voice and video, so as to guarantee QoS requirements for these real-time traffic. Note that the proposed scheme does not require any modification to the MAC protocol at wireless terminals. Through simulation experiments, we show the effectiveness of the proposed scheme.

Peer-to-Peer (P2P) file distribution systems can efficiently disseminate massive contents, such as disk images of operating systems, from a server to many users in a piece-by-piece manner. In particular, the BitTorrent protocol optimizes each peer's download speed by applying the tit-for-tat (TFT) strategy, where each peer preferentially uploads piece(s) to peer(s) from which it can download missing pieces faster. To the best of our knowledge, however, the optimality of TFT-based P2P file distribution has not been studied sufficiently. In this paper, we aim to understand the optimal scheduling in TFT-based P2P file distribution. First, we develop a discrete-time model of TFT-based P2P file distribution and formulate its optimal scheduling as a two-step integer linear programming problem. The first step is to minimize the average file retrieval time among peers, and the second step is to improve fairness among peers. We analyze the optimal solution obtained by the existing solver and reveal the characteristics of the optimal scheduling. Specifically, we show that it is crucial to distribute pieces from the server indirectly to peers with large upload capacity via those with small upload capacity.

As one of the technologies for the retrieval of desired contents over large-scale networks, multi-agent systems are receiving much attention. Since there are too many contents on the network to search them all exhaustively, some applications on multi-agent systems have time constraints, that is, they must obtain a result by a given deadline. To find better results for such applications, it is important for the agents to complete their tasks on as many nodes as possible by the deadline. However, most existing agent systems using round robin scheduling disciplines do not take time constraints into account. Therefore, agents are likely to miss their deadlines on many nodes. In this paper, we propose an efficient agent-dispatching method for time-constrained applications. This method decides creation and migration of a clone agent according to the estimated value of the number of agents that would have completed their tasks by the deadline. The results of our performance evaluation show that the proposed method increases the number of agents that complete their tasks.

Network tomography is an inference technique for internal network characteristics such as link loss rate and link delay from end-to-end measurements. In this paper, we consider network tomography for link loss rates, which is referred to as loss tomography. We propose a loss tomography scheme with bitwise operation-based in-network processing. Intermediate nodes generate coded packets by performing bitwise-operations on received packets so as to embed information about paths along which those packets have been transmitted. The coded packets are then forwarded to downstream nodes. In this way, receiver nodes obtain information about paths along which packets are transmitted successfully. Moreover, we show a recursion to compute the likelihood function of path loss rates, which can be utilized in estimating link loss rates from path loss information.

Mobile databases will play an important role in mobile computing environment, to provide data storing and data retrieval functionalities which are needed by most applications. In mobile computing environment, the wireless communication poses some problems, which require us to minimize its use. Replication is a database technique that is commonly used to fulfill the requirement in minimizing network usage. In this paper, we propose two replica allocation strategies, called primary-copy tracking replica allocation (PTRA) and user majority replica allocation (UMRA), which are better suited to the mobile computing environment. Their proposals are intended to cope with cost performance issues in data replication due to user mobility in mobile computing environment. To investigate their effectiveness, we provide access cost analysis and comparison on these strategies and the static replica allocation (SRA) strategy. We show that our proposed strategies outperform the SRA strategy when user mobility (inter-cell movement) is relatively low as compared with data access rate.

Communication among isolated networks (clusters) in delay tolerant networks (DTNs) can be supported by a message ferry, which collects bundles from clusters and delivers them to a sink node. When there are lots of distant static clusters, multiple message ferries and sink nodes will be required. In this paper, we aim to make groups, each of which consists of physically close clusters, a sink node, and a message ferry. Our objective is minimizing the overall mean delivery delay of bundles in consideration of both the offered load of clusters and distances between clusters and their sink nodes. Based on existing work, we first model this problem as a nonlinear integer programming. Using a commercial nonlinear solver, we obtain a quasi-optimal grouping. Through numerical evaluations, we show the fundamental characteristics of grouping, the impact of location limitation of base clusters, and the relationship between delivery delay and the number of base clusters.

We study wireless networked control systems (WNCSs), where controllers (CLs), controlled objects (COs), and other devices are connected through wireless networks. In WNCSs, COs can become unstable due to bursty packet losses and random delays on wireless networks. To reduce these network-induced effects, we utilize the packetized predictive control (PPC) method, where future control vectors to compensate bursty packet losses are generated in the receiving horizon manner, and they are packed into packets and transferred to a CO unit. In this paper, we extend the PPC method so as to compensate random delays as well as bursty packet losses. In the extended PPC method, generating many control vectors improves the robustness against both problems while it increases traffic on wireless networks. Therefore, we consider control vector selection to improve the robustness effectively under the constraint of single packet transmission. We first reconsider the input strategy of control vectors received by COs and propose a control vector selection scheme suitable for the strategy. In our selection scheme, control vectors are selected based on the estimated average and variance of round-trip delays. Moreover, we solve the problem that the CL may misconceive the CO's state due to insufficient information for state estimation. Simulation results show that our selection scheme achieves the higher robustness against both bursty packet losses and delays in terms of the 2-norm of the CO's state.

In this paper, we consider the broadcast storm problem in dense wireless ad hoc networks where interference among densely populated wireless nodes causes significant packet loss. To resolve the problem, we apply randomized network coding (RNC) to the networks. RNC is a completely different approach from existing techniques to resolve the problem, and it reduces the number of outstanding packets in the networks by encoding several packets into a single packet. RNC is a kind of linear network coding, and it is suited to wireless ad hoc networks because it can be implemented in a completely distributed manner. We describe a procedure for implementing the wireless ad hoc broadcasting with RNC. Further, with several simulation scenarios, we provide some insights on the relationship between the system parameters and performance and find that there is the optimal length of coding vectors for RNC in terms of packet loss probability. We also show a guideline for the parameter setting to resolve the broadcast storm problem successfully.

With the rapid increase of link speed in recent years, packet sampling has become a very attractive and scalable means in collecting flow statistics; however, it also makes inferring original flow characteristics much more difficult. In this paper, we develop techniques and schemes to identify flows with a very large number of packets (also known as heavy-hitter flows) from sampled flow statistics. Our approach follows a two-stage strategy: We first parametrically estimate the original flow length distribution from sampled flows. We then identify heavy-hitter flows with Bayes' theorem, where the flow length distribution estimated at the first stage is used as an a priori distribution. Our approach is validated and evaluated with publicly available packet traces. We show that our approach provides a very flexible framework in striking an appropriate balance between false positives and false negatives when sampling frequency is given.

We examine delay performance of packets from constant bit rate (CBR) traffic whose delay is affected by non-real-time traffic. The delay performance is analyzed by solving the Σ Di/G/1 queue with vacations. Our analysis allows heterogeneous service time and heterogeneous interarrival time. Thus, we can get the impact of packet length of a stream on the delay time of other streams. We then give various numerical results for enterprise multimedia networks, which include voice, video and data communication services. From our quantitative evaluation, we conclude that packet length of video traffic has large influence on the delay time of voice traffic while voice traffic gives a little impact on the delay time of video traffic.

Wireless networked control systems (WNCSs) are control systems whose components are connected through wireless networks. In WNCSs, a controlled object (CO) could become unstable due to bursty packet losses in addition to random packet losses and round-trip delays on wireless networks. In this paper, to reduce these network-induced effects, we propose a new design for multihop TDMA-based WNCSs with two-disjoint-path switching, where two disjoint paths are established between a controller and a CO, and they are switched if bursty packet losses are detected. In this system, we face the following two difficulties: (i) link scheduling in TDMA should be done in such a way that two paths can be switched without rescheduling, taking into account of the constraint of control systems. (ii) the conventional cross-layer design method of control systems is not directly applicable because round-trip delays may vary according to the path being used. Therefore, to overcome the difficulties raised by the two-path approach, we reformulate link scheduling in multihop TDMA and cross-layer design for control systems. Simulation results confirm that the proposed WNCS achieves better performance in terms of the 2-norm of CO's states.

Custody transfer in delay tolerant networks (DTNs) provides reliable end-to-end data delivery by delegating the responsibility of data transfer among special nodes (custodians) in a hop-by-hop manner. However, storage congestion occurs when data increases and/or the network is partitioned into multiple sub-networks for a long time. The storage congestion can be alleviated by message ferries which move around the network and proactively collect data from the custodians. In such a scenario, data should be aggregated to some custodians so that message ferries can collect them effectively. In this paper, we propose a scheme to aggregate data into selected custodians, called aggregators, in a fully distributed and autonomous manner with the help of evolutionary game theoretic approach. Through theoretical analysis and several simulation experiments, taking account of the uncooperative behavior of nodes, we show that aggregators can be selected in a self-organized manner and the number of aggregators can be controlled to a desired value.

Epidemic Routing is a data delivery scheme based on the store-carry-forward routing paradigm for sparsely populated mobile ad hoc networks. In Epidemic Routing, each node copies packets in its buffer into any other node that comes within its communication range. Although Epidemic Routing has short delay performance, it causes excessive buffer space utilization at nodes because many packet copies are disseminated over the network. In this paper, aiming at efficient buffer usage, we propose an XOR-based delivery scheme for Epidemic Routing, where nodes encode packets by XORing them when their buffers are full. Note that existing delivery schemes with coding are active coding, where source nodes always encode packets before transmitting them. On the other hand, the proposed scheme is passive coding, where source nodes encode packets only when buffer overflow would occur. Therefore, the behavior of the proposed scheme depends on the buffer utilization. More specifically, if sufficient buffer space is available, the proposed scheme delivers packets by the same operation as Epidemic Routing. Otherwise, it avoids buffer overflow by encoding packets. Simulation experiments show that the proposed scheme improves the packet delivery ratio.